November 10, 2006
@ 07:32 PM

Greg Linden has a blog post entitled Marissa Mayer at Web 2.0 where he writes

Marissa started with a story about a user test they did. They asked a group of Google searchers how many search results they wanted to see. Users asked for more, more than the ten results Google normally shows. More is more, they said.

So, Marissa ran an experiment where Google increased the number of search results to thirty. Traffic and revenue from Google searchers in the experimental group dropped by 20%.

Ouch. Why? Why, when users had asked for this, did they seem to hate it?

After a bit of looking, Marissa explained that they found an uncontrolled variable. The page with 10 results took .4 seconds to generate. The page with 30 results took .9 seconds.

Half a second delay caused a 20% drop in traffic. Half a second delay killed user satisfaction.

This conclusion may be surprising -- people notice a half second delay? -- but we had a similar experience at In A/B tests, we tried delaying the page in increments of 100 milliseconds and found that even very small delays would result in substantial and costly drops in revenue.

If you are a developing a consumer web site whose revenue depends on the number of page views you get, you need to print out that post and nail it to every bulletin board in your offices. One big problem with the AJAX craze that has hit the Web is how much slower websites have become now that using Flash and DHTML to add "richness" to Web applications is becoming more commonplace. My mind now boggles at the fact that I now see loading pages that last several seconds when visiting Web sites more and more these days.

Below is the graphic that shows up when you try to login to your Yahoo Mail beta inbox.

break dancing dude that shows up when you try to login to Yahoo Mail

My girlfriend watched me waiting for my inbox to show up while the above animated graphic was displaying and joked that they should change the graphic to a tortoise crawling across the screen so you have a heads up about how long it's going to take. :)

Of course, Windows Live services have also been guilty of this as are most of the 'Web 2.0' websites out there. At the end of the day, it's better for me to get to my data as quickly as possible than it is for the experience to be 'rich'. Remember that.


Categories: Web Development
Tracked by: [Pingback] [Pingback] [Pingback] [Pingback] [Pingback] [Pingback] [Pingback] [Pingback] [Pingback] [Pingback] [Pingback] [Pingback] [Pingback] [Pingback] [Pingback] [Pingback] [Pingback] [Pingback] [Pingback] [Pingback] [Pingback] [Pingback] [Pingback] [Pingback] [Pingback] [Pingback] [Pingback] [Pingback] [Pingback] [Pingback] [Pingback]

Friday, November 10, 2006 8:40:00 PM (GMT Standard Time, UTC+00:00)
But that's what boggles my mind ... why bother loading the HTML, then making *another* round trip to the server (or more) to actually request the data!?

web 2.0 apps should, when loading a page from anew, download a fully initialized UI, then make additional ajax requests as needed.
Saturday, November 11, 2006 1:47:58 AM (GMT Standard Time, UTC+00:00)
There's definitely a delicate balance in striking the right balance between using Ajax for the right purpose (avoiding page refreshes) and destroying a performant experience. We've been going through this as we work on the Version 4.0 of Channel 9 ( and are being conservative on our use of Ajax in this update. We tried to push the envelope on but found it wasn't always the best experience.
Saturday, November 11, 2006 7:01:46 AM (GMT Standard Time, UTC+00:00)
"There's definitely a delicate balance in striking the right balance between using Ajax for the right purpose (avoiding page refreshes) and destroying a performant experience."

A performant experience for who?

Instead, why not provide the choice to users? And, in parallel, pick by default the more performing UI based on the connection speed. I would guess someone connecting with dial-up (on Channel 9, that would be someone reading/commenting on forums, not watching videos) would prefer a thin experience. When your connection is slow, no wonder you prefer it barebone.

Also, the problem with testing mostly done with reliable broadband connections is that it disregards those connecting with unreliable connections. Ever experienced an Ajax application with with an unreliable connection?
Stephane Rodriguez
Monday, November 13, 2006 2:10:53 PM (GMT Standard Time, UTC+00:00)
Web can be fast too.

What you need is a hosts file that disabled all 3rd party ad-providers and use of Internet Explorer (No Add-ons).

That makes some sites even 5-10x faster than they normally open.

WPF/XBAP sites would open fast if the OS would warm up all those dlls needed to open them. Cold run can be like 10-30 seconds to open a simple site vs less than second warmed up.

If the first page load is 1 sec, it's ok IF I know there won't be more page loads.
Comments are closed.